metadata
dataset_info:
features:
- name: speechgeneration_q_qp_r_hidden_states
sequence:
sequence:
sequence: float32
splits:
- name: train
num_bytes: 782024644
num_examples: 1261
download_size: 782421537
dataset_size: 782024644
configs:
- config_name: default
data_files:
- split: train
path: data/train-*